A simplified method for raising antibodies to rat dopamine-beta-hydroxylase

Production and immunocytochemical application of a highly sensitive and specific monoclonal antibody against rat dopamine-beta-hydroxylase

A highly sensitive and specific monoclonal antibody against the enzyme dopamine beta-hydroxylase (DBH) from rat was produced and coded DBH 41. The generated hybridoma secreted immunoglobulins of mouse IgG1 subtype, as determined by radial immunodiffusion. This antibody, characterized by immunoblotting against a crude rat DBH preparation, was found to specifically recognize two bands of molecular weight 70 and 75 kDa corresponding to the soluble and membrane bound forms of the enzyme, respectively. With regard to species specificity, the anti-DBH antibody recognizes only the rat DBH molecule as it exhibits no cross-reactivity with either mouse, human, rabbit, guinea pig, cat or bovine DBH. Comparative immunocytochemical localization of DBH and TOH immunoreactivity was performed in different brain regions and we found that the DBH 41 antibody specifically stained DBH-containing neurons and fibers in the rat central nervous system (CNS). The high sensitivity of the DBH 41 antibody permitted us to detect immunologically the presence of the enzyme even in areas where only scattered DBH-containing fibers were present.

Transient tyrosine hydroxylase-like immunoreactive neurons contain somatostatin and substance P in the developing amygdala and bed nucleus of the stria terminalis of the rat

Tyrosine hydroxylase-like immunoreactive (TH-IR) neurons were observed from the embryonic day 17 (E17) to 6 weeks postnatally in two closely related nuclei of the limbic system, the bed nucleus of the stria terminalis (BNST) and the central nucleus of the amygdala (CNA) where they were restricted to circumscribed zones. These cells were scarce with an immature morphological aspect at E17. They progressively differentiated and increased in number until postnatal day 5 (P5), when their maximal density was reached. They were characterized as neurons by their ultrastructural appearance and the presence of both axo-somatic and axo-dendritic synaptic junctions. Moreover, TH-IR axons could be followed in the stria terminalis leaving the CNA, suggesting that part of TH-IR cells could be long projecting neurons rather than interneurons. A gradual decrease in the intensity of TH-IR and in density of labeled neurons was noted from P15 on, in both nuclei, (-50% at 4 weeks) until their total disappearance at 7 weeks. The significance of this TH-IR labeling regarding the catecholaminergic transmission remains unclear since these neurons did not contain the other catecholaminergic synthetic enzymes (DOPA-decarboxylase, dopamine-beta-hydroxylase, phenylethanolamine-N-methyl transferase) nor endogenous catecholamines. Double-labeling immunocytochemical methods, indicated that almost all the TH-IR neurons were colocalized with somatostatin 28 (SST) and with substance P (SP). Therefore these neurons expressed simultaneously 3 phenotypes, TH, SST and SP. This observation brings forth the notion of multiple neurotransmitter expression in transient neuronal populations and raises the question of neurotransmitter plasticity in the late postnatal development of the central nervous system (CNS). These neurons which were observed in two closely interconnected structures could be involved in early limbic circuits.

Postnatal sequential development of dopaminergic and enkephalinergic perineuronal formations in the lateral septal nucleus of the rat correlated with local neuronal maturation

Tyrosine-hydroxylase (TH-IR) and methionine-enkephalin like immunoreactivity (MetE-IR) were analyzed in the lateral septal nucleus (LSN) of the rat from birth (PO) to adulthood. TH-IR labeled specifically the dopaminergic (DA) pericellular arrangements of the LSN, as checked by negative dopamine-beta-hydroxylase and phenylethanolamine-N-methyl transferase-IR. TH-IR and Met-IR processes were present at birth in the medial LSN and extended lateralwards and caudalwards from P0 to P6 to constitute two main DA terminal fields (medial and lateral) surrounding a MetE one. Within these fields, the development of perineuronal baskets followed a similar medial to lateral sequence: DA axons first surrounded a few neuronal cell bodies at P3 in the medial part of the intermediate LSN; at P6, Met-IR axons encircled more laterally located perikarya, and only at P9, some neurons located along the ventricle in the lateral DA field became surrounded. The initial aspect of TH-IR baskets consisting of few axons surrounding the cell body rapidly evolved in a positive network encapsulating the perikaryon and long segments of the proximal dendrites, whereas MetE-IR varicosities remained restricted around the perikaryon and the initial dendritic segments. Ultrastructural study at P14 revealed numerous TH-IR and MetE-IR axosomatic and axodendritic profiles. TH-IR axosomatic varicosities exhibited asymmetrical synapses, whereas MetE-IR ones displayed rare symmetrical contacts. The medio-lateral gradient of development of the perineuronal baskets was parallel to the postnatal neuronal development of the LSN as evaluated by cytological criteria: neuronal density, cell size and Nissl staining. Therefore, the formation of DA and MetE perineuronal arrangements in the LSN does not seem to be subordinate to the nature of the neurotransmitter they contain but related to the level of differentiation of their target neurons. A similar sequential set-up in the development of afferences paralleling the neuronal differentiation is discussed.

Dopamine beta-hydroxylase: biochemistry and molecular biology

Dopamine beta-hydroxylase (DBH) catalyzes the conversion of dopamine to norepinephrine (NE), and is known to exist in two forms: soluble and membrane-bound. It has been reported that the two forms are similar in their immunoreactivity, carbohydrate content, and binding affinities for various substrates, and are apparently dissimilar in subunit structures and hydrophilicity. Furthermore, added structural complexity is observed within sDBH itself. Our results indicate that purified sDBH, which runs a single band on a nondenaturing gel, exhibits three protein bands of 75 kDa, 72 kDa, and 69 kDa on SDS polyacrylamide gel. The majority of sDBH exists as a 72-kDa protein. The NH2-terminal amino acid sequence of this 72-kDa protein indicates that it consists of two polypeptides of equimolar concentrations, where one differs from the other by three extra amino acids at its NH2 terminus. Whether they are different proteolytic cleavage products is not known. Thus, the structure of DBH appears to be more complex than originally considered. In vitro translation of total mRNA of bovine adrenal medulla followed by immunoprecipitation of DBH produces a single 72-kDa band on SDS polyacrylamide gel. This suggests either that there is only one in vitro mRNA translation product, which is modified to become different forms of DBH, or that multiple translation products are present but are indistinguishable by molecular weight. These subjects have been discussed in detail in this paper.

The distribution of neurotransmitter-specific cells and fibers in the anteroventral periventricular nucleus: implications for the control of gonadotropin secretion in the rat

The anteroventral periventricular nucleus (AVPv), which lies in the periventricular zone of the preoptic region, is critical for normal phasic gonadotropin secretion since lesions of this nucleus abolish the progesterone-induced surge of luteinizing hormone secretion from the anterior pituitary, block ovulation, and induce persistent vaginal estrus in female rats. However, very little is known about the neurotransmitter-specific pathways associated with this nucleus. In the present study we evaluated the distribution of biochemically specific cells and fibers within the AVPv and adjacent regions by using an indirect immunohistochemical method with antisera to serotonin (5-HT), dopamine beta-hydroxylase (DBH), tyrosine hydroxylase (TH), neuropeptide Y (NPY), cholecystokinin-8 (CCK), vasoactive intestinal polypeptide (VIP), substance P (SP), neurotensin (NT), corticotropin-releasing factor (CRF), luteotropin-releasing hormone (LRH), somatostatin (SS), thyrotropin-releasing hormone (TRH), oxytocin (OXY), vasopressin (VAS), adrenocorticotropic hormone (ACTH1-24), alpha-melanocyte-stimulating hormone (alpha-MSH), leucine-enkephalin (L-ENK), and calcitonin gene-related peptide (CGRP). Our findings indicate that both cells and fibers containing these putative neurotransmitters are differentially distributed in and around the AVPv in accordance with the cytoarchitectonic organization of this part of the preoptic region. The AVPv itself appears to receive strong inputs from SP-, VAS-, CCK-, and SS-containing pathways, whereas the highest densities of L-ENK-, NT-, 5-HT-, NPY-, and DBH-immunoreactive fibers were found in the cell-sparse zone just lateral to the AVPv. The suprachiasmatic preoptic nucleus (PSCh), a small group of cells located ventral to the AVPv just dorsal to the optic chiasm, contained high densities of alpha-MSH- and ACTH-immunoreactive fibers, as well as substantial numbers of fibers containing catecholamines or NPY. In contrast, a dense plexus of VAS-stained fibers was distributed fairly evenly throughout the AVPv and PSCh. Numerous L-ENK-immunoreactive cell bodies, and moderate numbers of CCK-, NT-, and CRF-stained cell bodies were found in the AVPv. The PSCh contained many TH-stained cells (presumably dopaminergic), in addition to a moderate number of CCK-containing cell bodies, while a high density of NT- and CRF-stained cells were found in the cell-sparse zone lateral to the AVPv, in addition to several CCK-, SP-, VIP-, and TH-containing cells.(ABSTRACT TRUNCATED AT 400 WORDS)

Neurotransmitter specificity of cells and fibers in the medial preoptic nucleus: an immunohistochemical study in the rat

The medial preoptic nucleus (MPN) is a sexually dimorphic complex with three major subdivisions. The cell-dense central (MPNc) and medial (MPNm) subdivisions are larger in male rats, while the cell-sparse lateral subdivision (MPNl) occupies a majority of the nucleus in females. In the present study we evaluated the distribution of possible monoaminergic and peptidergic cells and fibers within the MPN, as well as in adjacent regions of the medial preoptic area of the adult male rat. For this, we used an indirect immunohistochemical method with antisera to serotonin (5HT), dopamine beta-hydroxylase (DBH), tyrosine hydroxylase (TH), neuropeptide Y (NPY), cholecystokinin (CCK), vasoactive intestinal polypeptide (VIP), substance P (SP), neurotensin (NT), corticotropin-releasing factor (CRF), luteotropin-releasing hormone (LRH), somatostatin (SS), thyrotropin-releasing hormone (TRH), oxytocin (OXY), vasopressin (VAS), adrenocorticotropic hormone (1-24; ACTH), alpha-melanocyte-stimulating hormone (alpha-MSH), leucine-enkephalin (L-ENK), and calcitonin gene-related peptide (CGRP). The results suggest that cell bodies and/or fibers crossreacting with all of these putative neurotransmitters are differentially distributed within the MPN. Within the MPNm, the densest plexuses of fibers were stained with antisera to SP and NPY, while moderate densities of fibers were stained with anti-DBH, SS, CCK, CGRP, ACTH, and alpha-MSH, and only a few fibers were stained with anti-5HT, TH, NT, VAS, and L-ENK. Moderate numbers of SP- and L-ENK-immunoreactive cell bodies, and a few SS-, NT-, CRF-, and TRH-stained cell bodies were also found within the MPNm. The MPNc contained a dense plexus of CCK-immunoreactive fibers, as well as a few CRF-immunoreactive fibers. Both fiber types were localized almost exclusively to this subdivision, while most of the others studied here appeared to avoid it selectively. This suggests that there are relatively few inputs to the MPNc, and that they tend to avoid other parts of the nucleus, although moderate densities of DBH- and NPY-immunoreactive fibers were found in both the MPNm and MPNc. The MPNc contained several CCK-immunoreactive cell bodies as well as a moderate number of TRH-stained cell bodies. Both cell types were nearly completely localized to the MPNc. The major inputs to the MPNl studied here appear to be stained with antisera to 5HT and L-ENK, although moderate numbers of NT- and CRF- immunoreactive fibers were also found in this part of the nucleus.(ABSTRACT TRUNCATED AT 400 WORDS)

Transient expression of tyrosine hydroxylase immunoreactivity in some neurons of the rat neocortex during postnatal development

Tyrosine hydroxylase-like immunoreactive neurons were observed in the dorsolateral and medial neocortex of the rat during postnatal development. They occurred from 8 up to 24 days of age and lacked other catecholamine synthetizing enzymes. They appeared to be insensitive to the suppression of cortical noradrenergic innervation induced by neonatal subcutaneous injections of 6-hydroxydopamine.

New dopaminergic terminal fields in the motor, visual (area 18b) and retrosplenial cortex in the young and adult rat. Immunocytochemical and catecholamine histochemical analyses

New dopaminergic terminal fields have been visualized in the rat cerebral neocortex, using two morphological methods based on distinct properties of the dopaminergic system: presence of the first synthetic enzyme, tyrosine hydroxylase, and high-affinity uptake of amines. Tyrosine hydroxylase was used as an immunocytochemical marker after destruction of the cortical noradrenergic system, induced either neonatally by 6-hydroxydopamine or later on by DSP4, and controlled by the absence of dopamine beta-hydroxylase immunoreactivity. The uptake and storage of exogenous amines in tissue sections, in the presence of selective high-affinity transport inhibitors, enabled the specific visualization of the dopaminergic system with fluorescence histochemistry. A dopaminergic innervation of low density was observed along a dorsal sagittal strip which extended from the genu of corpus callosum until about 2 mm behind the splenium and encompassed several distinct cytoarchitectonic areas in the sensorimotor and visual cortex (medial and lateral agranular field, area 18b), as well as in discrete zones of the retrosplenial granular 29c,b, and agranular 29d areas. The distribution of these dopaminergic fields suggested a columnar organization. Several characteristics of the dopaminergic innervation were similar to that of the superficial anterior cingulate cortex (area 24): the laminar distribution to the superficial I-III layers, the secondarily developed varicose aspect in catecholamine fluorescence histochemistry and the delayed postnatal ingrowth in contrast with the early prenatal dopaminergic input to the prefrontal cortex. These similarities suggested that the subpopulation of dopaminergic neurons which provides projections to the anterior cingulate cortex could also contribute to the motor and visual cortex and thus play a role in sensorimotor integration. The predictive value of these results in the ascent of the phylogenetic scale are further considered.

Postnatal ontogenesis of the dopaminergic innervation in the rat anterior cingulate cortex (area 24). Immunocytochemical and catecholamine fluorescence histochemical analysis

The postnatal development of the dopaminergic input to the rat anterior cingulate cortex (area 24) was followed using anti-tyrosine hydroxylase immunocytochemistry and catecholamine fluorescence histochemistry in control and noradrenaline-depleted rats. Noradrenaline depletion in the cerebral cortex was obtained by peripheral injections of 6-hydroxydopamine (6-OHDA) at birth or N-2-chloroethyl-N-ethyl-2-bromobenzylamine (DSP4) at various postnatal ages and controlled by the absence of dopamine-beta-hydroxylase-labelled axons. The superficial and deep components of the anterior cingulate dopaminergic field developed at a different rate in control as well as lesioned rats. The deep supragenual dopaminergic field was already present at birth like the dopaminergic innervation of the prefrontal cortex area 32. In the superficial field, the molecular layer was reached first from postnatal day 3 (P3) on by positive axons running through the anterior hippocampal continuation and from P5-P6 on by another dopaminergic contingent coming through the deep dopaminergic field and giving off collaterals for layer III. The adult distribution pattern and striking varicose aspect were not reached until P21-P30 and a further increased density was observed until P60. The superficial cingulate dopaminergic field extended into the pregenual part of area 24b. The innervation of the superficial and deep layers of the rat anterior cingulate cortex by two distinct dopaminergic subpopulations, one of them closely related to that of prefrontal cortex area 32, could be compared with other laminar differences. The important functional implications of these data are further discussed.

Morphological evidence for a dopaminergic terminal field in the hippocampal formation of young and adult rat

We have visualized the dopaminergic innervation of the hippocampal formation of the rat using two morphological methods: (1) tyrosine hydroxylase immunocytochemistry on noradrenaline-depleted animals and (2) fluorescence histochemistry after the uptake and storage of dopamine on hippocampal slices in vitro. The noradrenergic hippocampal terminal fields were destroyed by neonatal neurotoxin pretreatment and the validity of the lesion checked by the absence of dopamine beta-hydroxylase immunoreactivity. As observed on early postnatal ages, dopaminergic axons reached the hippocampal formation through the fimbria and the alveus, but also through the supracallosal bundle and the ventral amygdaloid area-entorhinal cortex. The temporal (ventral and caudal) part of the hippocampal formation received the bulk of the dopaminergic innervation whereas no fibers were observed in the septal pole. Very few positive axons were visualized in the hilus of the gyrus dentatus and CA3 field, only near the temporal pole. CA1 field (stratum oriens) was innervated throughout its ventral part. The most innervated area was the ventral part--especially the deep layers--of the subiculum, in particular the prosubiculum. The dorsal part of the subiculum displayed some positive axons, although to a lesser extent. The pre- and parasubiculum contained a few positive axons. In addition, some immunoreactive axons were observed in the anterior hippocampal continuation and the indusium griseum. The ventral junction prosubiculum-CA1 field appears to be the main target area for the hippocampal dopaminergic innervation. It is interesting that the same areas are characterized by their projections to the nucleus accumbens which receives dopaminergic afferents. Thus, the hippocampostriatal projections, that represent a link between the limbic and central motor mechanisms, could be under dopaminergic influence.

The distribution of monoaminergic cells and fibers in a periventricular preoptic nucleus involved in the control of gonadotropin release: immunohistochemical evidence for a dopaminergic sexual dimorphism

A small, discrete nucleus at the rostral end of the third ventricle, the anteroventral periventricular nucleus (AVPv), has been reported to be involved in the control of gonadotropin release. Since monoaminergic neurotransmitter systems have also been implicated in this function we used an indirect immunohistochemical approach to examine the distribution of 3 monoaminergic neurotransmitter systems in this nucleus. Sections through the AVPv of both colchicine and non-colchicine-treated adult male and female Sprague-Dawley rats were processed for immunohistofluorescence with antisera directed against tyrosine hydroxylase (TH), dopamine beta-hydroxylase (DBH), or serotonin (5-HT), and were subsequently counterstained with the fluorescent Nissl stain ethidium bromide. The distributions of TH-, DBH- and 5-HT-immunoreactive neural elements within the AVPv were evaluated and a comparison was made between males and females. In both sexes, few 5-HT-stained fibers were seen within the borders of the AVPv, in contrast to the relatively high 5-HT-stained fiber density of the surrounding region. A dramatic sexual dimorphism was found in the distribution of TH-immunoreactive fibers and cell bodies. Compared to males, the AVPv in the female contained 3-4 times as many TH-stained perikarya, and a 2- to 3-fold greater density of TH-stained fibers. A low to moderate density of DBH-immunoreactive fibers, and no DBH-stained cell bodies, were seen in the nucleus. A clear sex difference was not found in the density of DBH-stained fibers in the AVPv, indicating that the sexual dimorphism in TH-immunoreactive neural elements in this nucleus is due to a greater density of dopaminergic fibers and a greater number of dopaminergic cell bodies in the female. These results suggest that dopamine may participate in the control of gonadotropin secretion at the level of the AVPv.

The organization of noradrenergic pathways from the brainstem to the paraventricular and supraoptic nuclei in the rat

Axonal transport and immunohistochemical methods have been used to clarify the organization of pathways from noradrenergic and adrenergic cell groups in the brainstem to the paraventricular (PVH) and supraoptic (SO) nuclei of the hypothalamus. First, the location of such cells was determined with a combined retrograde tracer-immunofluorescence method. The fluorescent tracer, True Blue, was injected into the PVH or the SO, and sections through the brainstem were stained with anti-(rat) DBH, a specific marker for noradrenergic and adrenergic neurons. It was found that, after injections in the PVH, doubly labeled neurons were confined almost exclusively to 3 cell groups, the A1 region of the ventral medulla, which contained a majority of such cells, the A2 region in the dorsal vagal complex, and the locus coeruleus (A6 region). After injections centered in the SO an even greater proportion of doubly labeled cells were found in the A1 region, although some were also found in the A2 and A6 regions. The topography of doubly labeled cells indicates that these projections arise primarily from noradrenergic neurons, although adrenergic cells in both the C1 and the C2 groups probably contribute as well. Because well over 80% of the retrogradely labeled cells in these three regions were also DBH-positive, we next placed injections of [3H]amino acids into each of them in different groups of animals, and traced the course and distribution of the ascending (presumably DBH-positive) projections to the PVH and SO in the resulting autoradiograms. Injections centered in the A1 region labeled a substantial projection to most parts of the parvocellular division of the PVH, and was most dense in the dorsal and medial parts. In addition, terminal fields were labeled on those parts of the magnocellular division of the PVH, and of the SO, in which vasopressinergic cell bodies are concentrated. Injections centered in the A2 region also labeled a projection to the parvocellular division of the PVH that was topographically similar, but less dense, than that from the A1 region. In contrast, [3H]amino acid injections centered in the locus coeruleus labeled a moderately dense projection to the PVH that was limited to the medialmost part of the parvocellular division. Neither the A2 nor the A6 cell groups project to the magnocellular parts of PVH, or to the SO. The autoradiographic material, and additional double-labeling experiments, were used to identify and to characterize projections that interconnect the A1, A2 and A6 regions, as well as possible projections from these cell groups to the spinal cord. These results may be summarized as follows: a substantial projection from the nucleus of the solitary tract to the A1 region was identified, but this pathway does not arise from catecholaminergic neurons in the A2 cell group. DBH-stained cells in the A1 region project back to the dorsal vagal complex, as well as quite massively to the locus coeruleus (A6 region)...

A method for tracing biochemically defined pathways in the central nervous system using combined fluorescence retrograde transport and immunohistochemical techniques

A simple method for the simultaneous localization of an antigen and a retrogradely transported fluorescent dye within single neurons is described. The method is based upon: (1) the efficiency of retrograde neuronal labeling with the fluorescent marker 'true blue'; (2) the near-quantitative persistence of retrogradely transported true blue localizations after subsequent processing of the tissue for immunohistochemistry; and (3) the ability to distinguish clearly between true blue- and immunohistochemically-stained cells by simply using appropriate excitation wavelengths for each. First we examined the characteristics of two fluorescent tracers which are effectively transported over long distances in the rat. The results confirm that true blue and bisbenzimide are transported from terminal fields to parent cell bodies and that both tracers are taken up and transported by damaged fibers and by undamaged fibers-of-passage. No evidence for transneuronal transport of either dye in the anterograde or in the retrograde direction was found. Next, using the projection of the paraventricular nucleus of the hypothalamus (PVH) to the spinal cord as a test system, it was found by direct cell counts that a considerably greater percentage of cells in a specific subdivision of the nucleus was labeled following true blue injections into the spinal cord (88%) than was labeled after comparable injections of bisbenzimide (58%), or horseradish peroxidase (HRP) (24%), or after HRP-polyacrylamide gel implants (39%). A comparison of cell counts of true blue-labeled cells in normal material and in series of adjacent sections that were processed for immunohistochemistry suggested that only about 5% of the true blue-labeled cells are no longer detectable following the immunohistochemical procedures employed. Finally, by coupling the fluorescent retrograde tracing method with an indirect immunofluorescence technique, we have been able to localize reproducibly both retrogradely transported true blue and an antigen in individual neurons. The perfusion and staining method employed provided adequate staining of cell bodies that cross-reacted with antisera to one or another of the 9 peptides and enzymes tested. The results indicate that true blue is a versatile and highly sensitive marker for retrograde tracing studies that can be used alone, or in conjunction with respect to their biochemistry as well as their efferent connections.

An immunohistochemical study of the organization of catecholaminergic cells and terminal fields in the paraventricular and supraoptic nuclei of the hypothalamus

The distribution of catecholaminergic fibers and cell bodies in the paraventricular and supraoptic nuclei of the hypothalamus was investigated with immunohistochemical methods in the adult albino rat. Sections through the nuclei were stained with antisera to the catecholamine synthesizing enzymes tyrosine hydroxylase (TH), dopamine-beta-hydroxylase (DBH), and phenylethanolamine-N-methyltransferase (PNMT). The results suggest that adrenergic (PNMT-stained) fibers innervate the entire parvocellular division of the paraventricular nucleus, although the highest density of fibers was found in the medial part of the division. Only widely scattered adrenergic fibers are found in the magnocellular division of the nucleus and in the supraoptic nucleus. Noradrenergic fibers appear to innervate the periventricular zone of the paraventricular nucleus and those parts of the paraventricular and supraoptic nuclei that contain predominantly vasopressinergic neurons in both the normal and in the homozygous Brattleboro rat. Significant numbers--somewhat more than 500--of dopaminergic (TH-stained) neurons are found in the paraventricular nucleus; the cells are distributed throughout the nucleus but are concentrated in the medial and periventricular parts of the parvocellular division. Double-labeling experiments with the retrogradely transported tracer true blue indicate that between 4% and 8% of the dopaminergic neurons in the paraventricular nucleus project to the region of the dorsal vagal complex and/or thoracic levels of the spinal cord. It is concluded that adrenergic inputs to the paraventricular nucleus may influence cells that project to the median eminence and to preganglionic autonomic cell groups in the medulla and spinal cord. Noradrenergic inputs to the supraoptic and paraventricular nuclei may influence primarily vasopressinergic cells that project to the posterior lobe of the pituitary, as well as cells in the periventricular part of the paraventricular nucleus that project to the median eminence.